Online Payment Fraud Detection Using Machine Learning

Online payment fraud has become a critical     challenge in the era of digital transactions, affecting businesses and customers globally. Traditional rule-based fraud detection systems often fail to adapt to the evolving nature of fraudulent activities. This study explores the application of machine learning techniques to detect online payment fraud effectively. This study investigates the application of machine learning algorithms, including Random Forest, k-Nearest Neighbors (KNN), Support Vector Machines (SVM), and Extreme Gradient Boosting (XGBoost), to effectively detect online payment fraud. The dataset used for this research consists of anonymized transaction records, characterized by significant class imbalance between fraudulent and legitimate transactions. Preprocessing steps include scaling, feature selection, and addressing data imbalance through Synthetic Minority Over-sampling Technique (SMOTE). Each algorithm is trained and evaluated using key metrics tailored for imbalanced datasets, such as precision, recall, F1-score, and the Area Under the Receiver Operating Characteristic Curve (AUC-ROC). Experimental results reveal that ensemble-based methods like Random Forest and XG Boost outperform KNN and SVM in both accuracy and computational efficiency, particularly in detecting minority (fraudulent) cases. While KNN and SVM demonstrate moderate performance, their scalability to large datasets is a challenge. XG Boost emerges as the most robust algorithm due to its ability to capture complex patterns with minimal false positives. This research concludes that the integration of advanced machine learning models like XG Boost into fraud detection pipelines can significantly enhance real-time detection capabilities, providing a scalable and reliable solution to mitigate financial losses caused by fraudulent activities.